How your answers becomea single number.

The question bank consists of 90 original questions, 30 of which come up each time you take the test. For each of the test's 30 slots, 3 variants were created from the same area and of the same difficulty – so every selection has the same structure and the results are comparable.

The number pair: questions per test / questions in the bank.

Matrix questions (Raven-style 3×3 grids of figures) are the purest measure of fluid intelligence, so they make up nearly half of the questions. Difficulty roughly increases as you go: each test has 8 easy, 12 medium and 10 hard questions, and the areas alternate.

Your browser remembers which questions you've already seen and favours unseen ones the next time – so across three attempts in a row, not a single question repeats.

Each correct answer earns points weighted by its difficulty; skipped and wrong answers score 0.

We divide the weighted score by the maximum achievable (giving a value between 0 and 1), then project it onto the usual IQ scale:

s  = weighted score / maximum
z  = (s − (0.54 + correction)) / 0.16
IQ = 100 + 15 · z        (clamped between 55 and 145)

The mean of 0.54 and the standard deviation of 0.16 describe the estimated population distribution for this question set. The percentile comes from the cumulative distribution function of the normal distribution: an IQ of 115, for example, is roughly the 84th percentile.

Fluid reasoning performance peaks in the mid-twenties and then declines slowly. That's why the results of older test-takers and those under 16 are measured against a slightly lower expected average.

• The measurement error of a 30-question online test is much larger than that of a 1–2-hour assessment administered by a psychologist. The result is best read as a band of ±8–10 points.
• The norm is estimated rather than based on a representative sample – so the exact figure is indicative.
• Retaking the test pushes the result upwards because of the learning effect.
• The result is not a diagnosis and is not suitable as a basis for educational, employment or medical decisions.

Taking the test is free; the detailed result is available with 7-day full access (€3.90; unless you cancel, €9.90/month from day 8, cancel anytime). Scoring happens on the server; the correct answers never reach your browser. When you pay, your answers are attached to the Stripe payment transaction in a short, encoded form, and the result page calculates your result from that – we don't store them in a separate database.

The state of an unfinished test is kept only by your own browser, so that you can continue. We don't ask for your name or a user account; card details are handled by Stripe and we never see them.